Why is Business Semantics such a Hot Topic?

An Interview with Dave McComb

Dave McComb and his team at Semantic Arts are a bunch of technically-deep software guys who’ve turned their attention to the subject of business semantics. Their first methodology is in the area of message-based integration, but they see a whole raft of applications for semantics-based solutions opening up in the near future. The next question I ask myself is whether we’re about to embark on another technology and buzzword bandwagon, or whether the semantics problem is really that deeply entrenched in our systems and we’re only just starting to realize it.

Wilshire Conferences (Wilshire): Good morning Dave. Let’s get right to the subject of “Semantics.” My trusty on-line dictionary defines “semantics” as “The study or science of meaning in language.” Is that a agreeable definition for our conversation today?

Dave McComb (McComb): Yeah, as long as you don’t ask “what’s the meaning of ‘meaning?’” Then it gets a bit deep.

Wilshire: So explain to me then what “Business Semantics” are, and why they are important.

McComb: It’s really just the meaning of the information we’ve been modeling all this time in our business applications. Right now there are two problems with the semantics in our systems, a short term problem and a longer term problem. The short term problem is that our lack of precision and meaning of the information in our systems leaves us with ambiguous data stores that we only partially understand. A little bit of analysis goes a long way with this. The longer term problem is that we are going to need our systems to help us with the interpretation of data, and unless we can come up with some better clues that what we have to date we’re going to get horrible results.

Wilshire: You’re writing a book on the subject of Business Semantics. Tell me what it will cover and why you think the time is right for it.

McComb: It’s meant to be the first book you’d get if you were interested in applying Semantics to Business Systems, and as such it covers the semantic angle on a wide range of sub disciplines. I’ve got the Semantic angle on Business Rules, on EAI, on data modeling, XML, Web Services as well as the more obvious Ontologies and the Semantic Web. In terms of why I think the time is right for this now, I think we’re rounding a corner now in the progression of Information Systems. I think we’ve at the point where we know how to solve all of what we thought were the “hard” problems (algorithms, performance, user interface) and now we left with the “really hard” problems of “what do we want our systems to do?” and “how can I get my systems to talk to each other without blowing my whole budget?”

Wilshire: I’ve got you working overtime at our upcoming Wilshire Meta-Data Conference and DAMA Symposium in Orlando at the end of April. Let’s talk first about the subject of “Implementing Message-Based Data Integration” on which you’re conducting a full-day tutorial with your colleague Simon Robe. Messaging is also a hot topic right now – can you tell me why it seems to have captured so much attention as an integration solution?

McComb: I think everyone has heard the economic argument enough times now that it is sinking in (you know the n2 v. n number of connectors argument). And people are seeing that messaging in general and XML messaging in particular seem to be pretty good ways of establishing an intermediary representation between systems, and thereby decouple them.

Wilshire: I understand there are different approaches to message-based integration? What approach will you be talking about during your tutorial?

McComb: I guess the biggest distinction is ad hoc versus planned. Most system integration is still pretty much ad hoc: do the integration for this project, we’ll figure out how to do the rest of the enterprise later. Also in the ad hoc camp is the idea that the API for any app that you currently have is the “given”, and the best you can do is map them to something else.

We’re taking more of an architectural and a planned approach to the problem. We say that what is core to a business is what we call a “message model” which is the semantics that will be shared between the applications in an enterprise. As long as each application can talk to the core, it needn’t worry about the other applications.

Wilshire: Why do you favor this approach versus alternative methods?

McComb: The promise of XML and messaging is to get loose coupling between systems, such that you could swap one out or make major changes without affecting the others. But without an architectural approach we think you’re going to recreate the situations that exist now, only with slightly better tools.

Wilshire: You’re also doing a conference session on Web Services. I think your view on web services is fairly neutral – you’re neither for nor against as a generalization. As such, what will you be talking about in Orlando?

McComb: We’ll get into some of the pitfalls to be avoided when using web services as an integration technology, what the current standards are missing, how we see the technology evolving, and the applicability of semantics to service discovery.

Wilshire: So it seems there are some strong relationships between these three subject areas – semantics, data integration and web services. I would think there are other emerging areas of technology where semantics expertise would be useful – search, business rules, content management and XML schema development. And of course we’ve all heard about the Semantic Web. Would you agree? And where do you think the highest payoffs are?

McComb: You shouldn’t have gotten me started…I see semantics everywhere and correspondingly the payoff is everywhere. Search and content and closely related, as they both deal with what is called “unstructured” data. There are two tricks to the problem, one is structuring the unstructured data, and at the moment that is mostly a manual process with some tool support from companies like Applied Semantics and Attunity. The other issue is once you’ve tagged it how do you find what you’re looking for and this is a job for Ontology based queries (a query that knows “Bordeaux” is a wine and to include references to Bordeaux when you asked for wine).

The basic idea of the Semantic Web is that searching through billions of documents on the web for what you want just doesn’t scale. We need some automated help. We also can’t rely on there being one big ontology that everyone subscribes to, we need to have a federation of bits of trusted knowledge that we use to do intelligent searches and to turn some of our routine activities over to software agents. This is a long term project, but it’s worth tracking because of what it’s likely to throw off along the way.

Wilshire: For some time now I’ve suggested that data managers are ideally suited to move into the new roles being created by trends such as semantics, web services and XML schema development. Can you validate my thinking on this? It seems that the business and technical skills associated with metadata management are the same ones required in these emerging disciplines.

McComb: Lots of data administrators and modelers have been pretty bored lately: companies are buying packages and making incremental changes to legacy systems so there isn’t nearly as much modeling going on in corporate America. So one of the things we’re doing with our clients now is “repurposing” some of the DA’s into a role we call “Enterprise Message Modelers” they have the right background, and the organization is already in place to have a shared resource with some clout. It does seem that managing Web Services may fall into this same area, but we haven’t gone there yet.

Wilshire: So, what else do you do Dave? Your web site talks about “Semantic Driven Legacy Mining” and “Greenfield Semantic Design”. Impressive marketing names to be sure, but what on earth do they mean?

McComb: Mostly they mean we haven’t updated our web site in over a year, maybe we’ll have that taken care of by the time this gets out. The short answer to what we do is we help people with what we call “architectural remodeling,” and semantically inspired design. Most companies have accreted a collection of technologies that are dis-economically suiting their purposes. We use messaging-based approaches to help them work out strategies to improve them incrementally. And our messaging strategies, as well as all our design work is based on semantics, getting to the core meaning of the information makes sure that among other things you’re normalizing the right data.

Wilshire: When does the book come out?

McComb: The end of June .

Wilshire: Thanks Dave. You’ve given us lots to chew on here. Good luck with the new book.

Join us for the
Wilshire Meta-Data Conference
and DAMA International Symposium
May 2-6, 2004 • Century Plaza Hotel • Los Angeles, California USA

The World's Largest Vendor-Neutral Data Management Conference

The 16th annual DAMA International Symposium and 8th annual Wilshire Meta-Data Conference will be held May 2-6, 2004 at the Century Plaza Hotel in Los Angeles, a beautiful venue adjacent to Beverly Hills. Hear 40 case studies outlining strategies of companies that have implemented successful data management projects. There will be more than 120 speakers in all, covering meta data, enterprise architecture, data and process modeling, unstructured data, business rules, data integration, XML, business intelligence, data warehousing, information stewardship, and more. Keynote Speaker Chris Date. Click here for details.

©2003 Wilshire Conferences, Inc. May be quoted with full attribution.